© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 27111
    Prev
    Next
    INDEX
    Explanations

    specific legal terms and conditions related to disclaimers and rights

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    12
    +0.19
    1.1%
    248
    +0.14
    0.8%
    357
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    248
    +0.19
    0.13
    12
    +0.14
    0.13
    420
    +0.10
    0.08
    Negative Logits
    Ļª
    -3.07
    į
    -3.00
    ĭ
    -2.84
    ©
    -2.79
    ĩ
    -2.73
    Ģ
    -2.71
    Ĵ
    -2.71
    §
    -2.71
    ¿½
    -2.69
    ī
    -2.59
    POSITIVE LOGITS
     branch
    1.73
     <
    1.69
     attendees
    1.55
     organizers
    1.49
     suite
    1.44
     $
    1.43
     $_
    1.42
     Edition
    1.41
     Front
    1.40
     alumni
    1.38
    Activations Density 4.123%

    No Known Activations